SWEMSA 2019 | https://doi.org/10.5281/zenodo.3499650

LC-MS data analysis with xcms

  • xcms: toolbox for LC-MS data analysis.
  • Chromatographic peak detection: findChromPeaks.

  • Alignment: adjustRtime.

  • Correspondence: groupChromPreaks.

  • Result: matrix with feature abundances in samples.
  • Normalization and identification of features of interest: use any functionality available in R.
  • Annotation of features from LC-MS experiments challenging.
  • LC-MS/MS data: MS2 spectra assist in annotation.
  • Added support for LC-MS/MS data analysis in xcms.
  • Possible because we re-use functionality and classes from the MSnbase package.

Analyzing DDA data with xcms

dda_file <- system.file("TripleTOF-SWATH/PestMix1_DDA.mzML",
                        package = "msdata")
dda_data <- readMSData(dda_file, mode = "onDisk")
table(msLevel(dda_data))
## 
##    1    2 
## 4627 2975

Analyzing DDA data with xcms

  • MS1 chromatographic peak detection:
cwp <- CentWaveParam(snthresh = 5, noise = 100, ppm = 10,
                     peakwidth = c(3, 30))
dda_data <- findChromPeaks(dda_data, param = cwp)
dda_spectra <- chromPeakSpectra(dda_data)
dda_spectra
## Spectra with 158 spectra and 1 metadata column(s):
##                   msLevel     rtime peaksCount |     peak_id
##                 <integer> <numeric>  <integer> | <character>
##   CP01.F1.S1000         2   128.237         16 |        CP01
##   CP01.F1.S1008         2   128.737         40 |        CP01
##             ...       ...       ...        ... .         ...
##   CP98.F1.S5266         2   596.054         88 |        CP98
##   CP99.F1.S7344         2   873.714         20 |        CP99

Analyzing DDA data with xcms

  • Example: annotate chrom peak with an m/z of 304.1131.
chromPeaks(dda_data, mz = 304.1131, ppm = 20)
##            mz    mzmin    mzmax      rt   rtmin   rtmax    into     intb
## CP53 304.1133 304.1126 304.1143 424.614 417.985 430.784 13709.7 13658.01
##          maxo sn sample
## CP53 3978.987 74      1
  • Get MS2 spectra associated with that peak
ex_spectra <- dda_spectra[mcols(dda_spectra)$peak_id == "CP53"]

Analyzing DDA data with xcms

## Spectra with 5 spectra and 1 metadata column(s):
##                   msLevel     rtime peaksCount |     peak_id
##                 <integer> <numeric>  <integer> | <character>
##   CP53.F1.S3505         2   418.926         10 |        CP53
##   CP53.F1.S3510         2   419.306         30 |        CP53
##   CP53.F1.S3582         2   423.036        694 |        CP53
##   CP53.F1.S3603         2   423.966        783 |        CP53
##   CP53.F1.S3609         2   424.296        753 |        CP53
  • Build consensus spectrum.
ex_spectrum <- combineSpectra(ex_spectra, method = consensusSpectrum,
                              ppm = 10, minProp = 0.8)
ex_spectrum
## Spectra with 1 spectra and 1 metadata column(s):
##                   msLevel     rtime peaksCount |     peak_id
##                 <integer> <numeric>  <integer> | <character>
##   CP53.F1.S3505         2   418.926         17 |        CP53

Analyzing DDA data with xcms

  • Compare the consensus spectrum against 2 candidates with same m/z.
par(mfrow = c(1, 2))
plot(ex_spectrum, flumanezil, main = "Flumanezil", tolerance = 40e-6)
plot(ex_spectrum, fenamiphos, main = "Fenamiphos", tolerance = 40e-6)

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

SWATH data

swath_data <- readMSData("PestMix1_SWATH.mzML", mode = "onDisk")

Analyzing SWATH data with xcms

  • Chromatographic peak detection in MS1.
cwp <- CentWaveParam(snthresh = 5, noise = 100, ppm = 10,
                     peakwidth = c(3, 30))
swath_data <- findChromPeaks(swath_data, param = cwp)

Analyzing SWATH data with xcms

  • Chromatographic peak detection in MS1.
cwp <- CentWaveParam(snthresh = 5, noise = 100, ppm = 10,
                     peakwidth = c(3, 30))
swath_data <- findChromPeaks(swath_data, param = cwp)
  • Chromatographic peak detection in MS2 (within each isolation window).
swath_data <- findChromPeaksIsolationWindow(swath_data, param = cwp)

Analyzing SWATH data with xcms

  • Chromatographic peak detection in MS1.
cwp <- CentWaveParam(snthresh = 5, noise = 100, ppm = 10,
                     peakwidth = c(3, 30))
swath_data <- findChromPeaks(swath_data, param = cwp)
  • Chromatographic peak detection in MS2 (within each isolation window).
swath_data <- findChromPeaksIsolationWindow(swath_data, param = cwp)

Analyzing SWATH data with xcms

  • Reconstructing MS2 spectrum from SWATH data:
swath_spectra <- reconstructChromPeakSpectra(swath_data, minCor = 0.9)

  • For each MS1 chromatographic peak:

    • Find MS2 peaks (within the correct isolation window) with similar retention time.
    • Correlate peak shape of MS1 and candidate MS2 peaks.
    • Reconstruct the MS2 spectra based on matching MS2 peaks’ m/z and intensity.

Analyzing SWATH data with xcms

  • Example: reconstructed MS2 spectrum for Fenamiphos.
chromPeaks(swath_data, mz = 304.1131, ppm = 20, msLevel = 1L)
##            mz    mzmin    mzmax      rt   rtmin   rtmax     into     intb
## CP35 304.1124 304.1121 304.1126 423.945 419.445 428.444 10697.34 10688.34
##          maxo  sn sample
## CP35 2401.849 618      1
swath_sp <- swath_spectra[mcols(swath_spectra)$peak_id == "CP35"]
swath_sp
## Spectra with 1 spectra and 3 metadata column(s):
##       msLevel     rtime peaksCount |           ms2_peak_id
##     <integer> <numeric>  <integer> |       <CharacterList>
##   1         2        NA         15 | CP205,CP207,CP217,...
##                                                  ms2_peak_cor     peak_id
##                                                 <NumericList> <character>
##   1 0.999787099100833,0.964862731008839,0.980288982893133,...        CP35

Analyzing SWATH data with xcms

par(mfrow = c(1, 2))
plot(swath_sp[[1]], ex_spectrum, main = "DDA", tolerance = 40e-6)
plot(swath_sp[[1]], fenamiphos, main = "Fenamiphos", tolerance = 40e-6)

Annotating MS2 spectra

  • Compare Spectra against reference Spectra:
    • import from mgf file(s).
    • … (future developments…)

Future developments

  • Define an efficient, flexible and well documented infrastructure for Mass Spectrometry data in R.

  • Provide core functionality.
  • Provide core data representations.
  • Reusable in other packages.

  • Introduction of backends: independence between MS functionality and data origin/storage.

  • Introduction of backends: independence between MS functionality and data origin/storage.

  • Introduction of backends: independence between MS functionality and data origin/storage.

  • Introduction of backends: independence between MS functionality and data origin/storage.

  • Introduction of backends: independence between MS functionality and data origin/storage.

Thank you for your attention

Also thanks to: Micheal Witting, Jan Stanstrup, Steffen Neumann, Sebastian Gibb, Laurent Gatto